Siliceous and Delicious

Wednesday

Spent the afternoon on a long phone call with Zoe, who wanted to know more about age models and how to construct them than I was able to tell her. And then she wanted my feedback on her silicon isotope ideas (she’s made measurements of the silicon isotope composition of diatoms across the Cenozoic). It was a little frustrating, because I see some major problems with her interpretation, which is essentially also the canonical model that’s been applied by others who have looked at diatom silicon isotopes. And I just don’t understand it. Oh well!

I used the remaining time to prepare a couple more SEM stubs so I have something to sputter coat during my training tomorrow.

Thursday

Another pleasant morning at Darwin’s, my new second home. Continued my gastropodian sidle through Rabosky’s code. Marginal progress: I think I’ve managed to carry out my first subsampling analysis. It’s the simplest kind (naive rarefaction), and it doesn’t apply any of the “sampling probability” corrections that Rabosky adds in. In fact, he doesn’t even show this plot in his paper. Anyway, it gives the same result (qualitatively) that his other analyses give, too: the peak diversity is at around 30 Ma. Here is the plot for 50 iterations, with each time bin subsampled to 100 occurrences:

This is kind of an interesting plot, actually, because I think (but I haven’t thought this through in detail) it does not suffer from the same sort of biogeographic bias that the subsampling by-list algorithm has. Which was one of the main avenues of criticism I had been talking about with Charles.

The afternoon consisted of two activities. The first was the latest (and hopefully final) chapter of the Australia Reimbursement Fiasco. I finally got the reimbursement check from Harvard (in record time—barely two months after submitting my paperwork) and took it to the bank to have it turned into a bank check for Jc, who had written me a check to cover my credit card bill while I was waiting for Harvard to repay me. It was only when I got back to the office that I realized the check I had gotten from Harvard was for both Jc’s rental car expense and my personal expenses in Australia (hotel, meals, etc), so I had to go back and redeposit the check and have a new one issued for the right amount… Anyway. I’m hopeful that this saga has reached its denouement.

The second part of the afternoon was spent in the cavernous depths of the Center for Nanoscale Systems, four stories beneath Oxford Street, learning how to use the FIB-SEM in my second training session. I had taken the sputter coater certification course (which lasted a glorious ten minutes) so I had a sample ready to go. Lo and behold, when I had loaded it in the machine, it was spattered with countless little diatom frustules… success! Ah, the joy of seeing something tangible (at least theoretically speaking, ignoring the fact the damn things are hundredths of a millimeter across) to work on. Even better, the afternoon’s task consisted of obliterating one of them with a high-energy beam of ionized gallium. Very, very cool. Step one: find the poor sucker who’s going to get it. There. Set up the target for firing:

Step two: set phasers to kill, blast away!

Haha. Eat gallium and die, atom. As you can see the high-current beam has taken off the required chunk of diatom (and the carbon tape it’s glued to, as well!), but it’s left quite a ragged edge. So step three is to define a smaller target (see box in image) and fire a lower-current beam to get a smoother cut of the cross-section edge. Fire!

Almost needless to say, this was rather an enjoyable experience. There’s nothing quite like playing with multi-million dollar space-age equipment you’re absolutely unqualified to operate… In spite of this exhilaration, a couple of problems immediately come to mind. The first: as you can see from the last two images, it is quite hard to tell where the diatom frustule ends and the carbon tape begins at the tangent, where frustule sits on the tape. I’m not sure whether it’ll be possible, and if so how hard it will be, to resolve that boundary—which I will need to do if I am to measure the cross-sectional area of the cut surface. The root of the problem, I think, is the way the SEM makes images: it fires electrons at the sample, and records the lower-energy secondary electrons that are emitted by the sample as a result of this bombardment. This basically gives you the outline shape of the specimen’s surface but essentially no information about composition; hence the smooth edge of the cut surface in the last image above looks the same above (cut silica) and below (cut carbon). Not sure how to get around this.

The second problem that I will need to think about has to do with angles. The electron beam and the focused ion beam are offset at an angle of 54˚, and the stage will tilt to a maximum angle of 63˚ from horizontal. While this means you can cut a surface vertically into the sample (as in the example above), you can only view that surface at an oblique angle, at best 63˚ from the cut surface (again as in the example above—you will notice the view of the cut surface is not end-on, but at an angle from above). This is also not ideal for making measurements of area, as the angle both introduces distortion and may obscure some features from view. Possible solutions include mounting the sample on stubs that have angled surfaces and transferring the FIBbed sample to a different-angled stub. I need to think about both of those further.

Friday

Another day, another fresh start. It’s cold, crisp, and clear. Sunny, bright, and the wind is tugging at the last few dry, brown leaves clinging to the trees outside of Darwin’s. Why go to the office and surround myself with busy-body overachievers when there’s a cozy chair and a warm cup of coffee waiting for me here, where I can slouch among fellow layabouts and crazies, and watch the world go by outside the big windows.

Spent the first part of the morning keeping up with email, responding to Lars (Christian’s student) who had asked about how the FIBbing was going. And finally sent that email to Dave to ask his sage opinion on the feasibility of my silicon isotope study idea.

Monday

Hmm. Today might just as easily not have happened. I think, perhaps, the time has come to return to a slightly more rigid schedule—I’m starting to feel adrift, and lost, and unmotivated. Which, with the exception of the last adjective, weren’t the bulk of my problems before. Anyway, I’ve been feeling a bit run down all day, and unfocused as a result. Spent some time in the lab, mostly being frustrated by the fact that none of my radiolarian samples appear to be dissolving at all, in spite of them being in really very strong acid. If this was more of a lab of my ideal world, I’d ask someone for help, but as it stands I just feel stuck and lost. Ben (the other Ben, one of our new postdocs) would be a good start, but he’s away at GSA. Yech.

Had a short conversation with Andy over coffee—I told him about having completed my training on the FIB, and about the possibility of using the tool for serial-sectioning. His response was that I had found a really expensive way to make optical sections of a diatom, and suggested I try to make my reconstructions optically instead. Aaargh! What? I don’t know that Andy has any idea at all at this stage of what I am doing, or why I am doing it, and that scares me. Amidst all the brew-ha around the time of my quals, he was the only one who stood behind me and kept telling me that I was doing the right thing… what if all of a sudden he joins the ranks of the “what-you’re-doing-is-daft”-sayers? Since I am their president and spokesperson, that leaves nobody in the whole world to speak for my project, and isn’t a rosy outlook for my thesis or graduation. Ick.

Anyway. Thought a little bit about the angle issue with the FIB, and I think I’ve figured it out (famous last words). Basically, I think using the stubs Nicholas showed me, which have a section of their platform tilted at a 45˚ angle, should work. The FIB column is inclined by 54˚ relative to the zero-tilt of the stage, so during “normal” FIB operation you tilt the stage to 54˚. So at a tilt angle of 9˚ (54-45) I should be able to get the same effect using the angled stub (see drawing on left). Except that I could then further tilt the stage to 45˚, which (adding in the 45˚ of the stub itself) would give me a 90˚ view of the specimen, i.e. straight down onto the cross section I will have just FIBbed (see drawing on right). Which is an improvement over the 63˚ maximum afforded by the tilt of the stage itself.

Tuesday, Wednesday, Thursday

Sick days.

Friday

Went to the clinic this morning as my health wasn’t improving substantially. Apparently everything is in good order and it seems to be nothing worse than a persistent chest cold, though it seems to have caused an asthmatic response (hence the shortness of breath and chest tightness).

Trundled into the office not long before lunch and spent time filtering the one and only radiolarian sample so far that has actually dissolved. The residue I am left with doesn’t look promising—rather than a fine white powder (as I might have anticipated) it’s a thick mass of dark brown sludge. Yech. If there are any radiolarians in it, it’s going to be an arse trying to find them. In the afternoon, decided to make a polished section of a stromatolite hand sample I’d brought back from Australia; once upon a time it was going to be a gift for Kati. It doesn’t look as awesome as I had hoped. Ah well. Andy has called us for a holiday lab meeting in five minutes, so given that there’s only Monday left before I leave for Christmas, I think this is the point at which I will slam the covers shut on this academic year and look forward with hope and optimism to the year ahead. May it be infinitely more enjoyable and productive than this one has been.